World record for AMD

Toggle sidebar Toggle sidebar
V

Vegitto

Diamond Member
May 3, 2005
5,234
1
0
Originally posted by: SrGuapo
Originally posted by: Intelia
No just thought you might find it interesting.
Click to expand...

Indeed it is. Lets not crap on a perfectly good thread vegitto.
Click to expand...

Okay, okay. It's cool, I wish I could get my hands on one of those.
 
Elfear

Elfear

Diamond Member
May 30, 2004
7,163
819
126
Very nice results! Wish I had the money for one of those bad boys.

Intelia- Appreciate that even though you bat for Intel you're willing to post up accomplishments about AMD.
 
V

Valkerie

Banned
May 28, 2005
1,148
0
0
Intel or AMD, I don't care, whatevers faster and cheaper. It's pointless to assume sides, it's not like you own the company or anything.

Anyhow, on an FX, that is terribly sickening for performance. If I had the money, I'd probably go for that rig and all it's specs...
 
I

Intelia

Banned
May 12, 2005
832
0
0
You see it as baiting because of your views. Its not so at all. You don't know me personally so your assuming. Its alright though as long as we keep it clean I don't mind.

Now that Intel may be going to ondie memory controller in the future I may change my mind.

I really like the advantages of the contoller being on the chipset. Even though there is a large proformance hit .

But now with PCI-E going ondie I don't know I well wait and see if the same disadvantages are their as with ondie memory controller .

The thing that really gets me is you guys are all smart enough to now there's a hugh advantage to the the ondie memory controller and because of this AMD has an egde in performance.

Now it seems that AMD'S advantage here is having problems scaling up to higher memory speeds. So now Intel is going that route. Sometimes I get so confused.
Maybe intels cpus won't suffer this fate with the new Pressies but I am thinking the P3 arch. well also run into this wall.
So now Iam beginning to think Intel pressies may not be as bad as I precieved them to be . Man around and around we go. I well know much more with the release of the new Intel cpu's on 65 nano.chips if Intel gets the power usage and heat under controll netburst may still be alive Don't say it I know I know Ok! Lets say we can get to DDR2 to 1600 at low lati. And pressies can use that speed to its advantage. Who knows? Than indeed the ondie memory contoller well benefit the Pressies more than it will the conroe.

Confused yet I sure as hell am.
 
L

Lithan

Platinum Member
Aug 2, 2004
2,919
0
0
Yeah. It's a good run, but AMD really can't compete in SPI. Intel is just too strong there. Occassionally a crazy amd rig will take down the top p4... but nothing can touch these dothans in Spi. Even a dothan on air can often times beat Phase change cooled amd's and p4's.


Amd really has no problems scaling memory. They aren't doing it largely because DDR1 is dead and they are busy working on their DDR2 implimentation now. What they do/will have problems with (as will intel) is driving the memory demands of multicore rigs, especially as we move from 1gig in most systems to 2 gig, 4 gig, 16 gig, etc. I wouldn't be surprised if within two years amd has announced multicores with independant memory controllers. Not sure about intel though. If their debut "on-die" memory controller dual cores dont have it, I suspect they will come at least a year after amd with them.
 
I

imported_michaelpatrick33

Platinum Member
Jun 19, 2004
2,364
0
0
Originally posted by: Intelia
You see it as baiting because of your views. Its not so at all. You don't know me personally so your assuming. Its alright though as long as we keep it clean I don't mind.

Now that Intel may be going to ondie memory controller in the future I may change my mind.

I really like the advantages of the contoller being on the chipset. Even though there is a large proformance hit .

But now with PCI-E going ondie I don't know I well wait and see if the same disadvantages are their as with ondie memory controller .

The thing that really gets me is you guys are all smart enough to now there's a hugh advantage to the the ondie memory controller and because of this AMD has an egde in performance.

Now it seems that AMD'S advantage here is having problems scaling up to higher memory speeds. So now Intel is going that route. Sometimes I get so confused.
Maybe intels cpus won't suffer this fate with the new Pressies but I am thinking the P3 arch. well also run into this wall.
So now Iam beginning to think Intel pressies may not be as bad as I precieved them to be . Man around and around we go. I well know much more with the release of the new Intel cpu's on 65 nano.chips if Intel gets the power usage and heat under controll netburst may still be alive Don't say it I know I know Ok! Lets say we can get to DDR2 to 1600 at low lati. And pressies can use that speed to its advantage. Who knows? Than indeed the ondie memory contoller well benefit the Pressies more than it will the conroe.

Confused yet I sure as hell am.
Click to expand...

AMD isn't having any problems scaling up in memory support. The JEDEC standard for DDR1 is DDR400 (PC3200 etc.). Any higher memory frequency would be non standard and AMD isn't going to officially release a processor with non-standard memory as a base. (Hidden overclocker options are just that, options)

DDR2 to 1600 is not going to happen. DDR 2 to 1000 maybe. DDR3 1600 isn't even on graphcis cards much less main system memory. AMD at least in the next revision, is going to DDR2 667 on die. Until the latencies of DDR2 667 and DDR 2 800 came (is still coming down) it was pointless for AMD to upgrade. There really isn't much difference in performance bewteen DDR 667 with higher latencies and DDR 400 with lower latencies. Now that DDR latencies and prices are coming down we notice AMD having the DDR 2 (and 3) cpus coming to market.

The world record is meant to be broken. 3.66 FX (MMMM drools)
 
I

Intelia

Banned
May 12, 2005
832
0
0
Intels ondie controller well implament a token ring type design so they may bypass these problem with a small lati. hit. Just guesing but I think maybe Intels engineers caught this. Time well tell. Amd processors really don't seem to benefit much from DDr 500 at low lati. so their does seem to be a head here maybe its in the controller but I am thinking not. The easiest way for me to explain what I am seeing . would be to use a closed hydro loop. The Cpu being the motor and the memory being the pump. So the water being pumped would be the information. As I see it Amd has a more effient pump so its system runs more effiently. SO now with a higher pressure pump(memory) it won't let the motor (cpu) run at higher pressure(the information) because the motor can't handle the higher pressure. I know thats over simplifieng but thats how it infact does work. It really doesn't matter if your talking about water ,air or electicity flow is flow . so if you run into a head pressure problem the whole system has to change .(In this case it seems to be the cpu is the bottle neck and that is showing in the 7800 gpu test.). pump ,motor and piping to reduce head pressure. and gain more flow at the same time.

This is were the Intel pressies may have the advantage longer pipes(BAD thing ) More pipes (good thing)
 
I

Intelia

Banned
May 12, 2005
832
0
0
Originally posted by: michaelpatrick33
Originally posted by: Intelia
You see it as baiting because of your views. Its not so at all. You don't know me personally so your assuming. Its alright though as long as we keep it clean I don't mind.

Now that Intel may be going to ondie memory controller in the future I may change my mind.

I really like the advantages of the contoller being on the chipset. Even though there is a large proformance hit .

But now with PCI-E going ondie I don't know I well wait and see if the same disadvantages are their as with ondie memory controller .

The thing that really gets me is you guys are all smart enough to now there's a hugh advantage to the the ondie memory controller and because of this AMD has an egde in performance.

Now it seems that AMD'S advantage here is having problems scaling up to higher memory speeds. So now Intel is going that route. Sometimes I get so confused.
Maybe intels cpus won't suffer this fate with the new Pressies but I am thinking the P3 arch. well also run into this wall.
So now Iam beginning to think Intel pressies may not be as bad as I precieved them to be . Man around and around we go. I well know much more with the release of the new Intel cpu's on 65 nano.chips if Intel gets the power usage and heat under controll netburst may still be alive Don't say it I know I know Ok! Lets say we can get to DDR2 to 1600 at low lati. And pressies can use that speed to its advantage. Who knows? Than indeed the ondie memory contoller well benefit the Pressies more than it will the conroe.

Confused yet I sure as hell am.
Click to expand...

AMD isn't having any problems scaling up in memory support. The JEDEC standard for DDR1 is DDR400 (PC3200 etc.). Any higher memory frequency would be non standard and AMD isn't going to officially release a processor with non-standard memory as a base. (Hidden overclocker options are just that, options)

DDR2 to 1600 is not going to happen. DDR 2 to 1000 maybe. DDR3 1600 isn't even on graphcis cards much less main system memory. AMD at least in the next revision, is going to DDR2 667 on die. Until the latencies of DDR2 667 and DDR 2 800 came (is still coming down) it was pointless for AMD to upgrade. There really isn't much difference in performance bewteen DDR 667 with higher latencies and DDR 400 with lower latencies. Now that DDR latencies and prices are coming down we notice AMD having the DDR 2 (and 3) cpus coming to market.

The world record is meant to be broken. 3.66 FX (MMMM drools)
Click to expand...

I was referring to AMDs new 500 standard and their does indeed seem to be a problem here. As some here know I am sure of that. I am not new to this so I will try to see it apples to apples ok
 
I

imported_michaelpatrick33

Platinum Member
Jun 19, 2004
2,364
0
0
Originally posted by: Intelia
Originally posted by: michaelpatrick33
Originally posted by: Intelia
You see it as baiting because of your views. Its not so at all. You don't know me personally so your assuming. Its alright though as long as we keep it clean I don't mind.

Now that Intel may be going to ondie memory controller in the future I may change my mind.

I really like the advantages of the contoller being on the chipset. Even though there is a large proformance hit .

But now with PCI-E going ondie I don't know I well wait and see if the same disadvantages are their as with ondie memory controller .

The thing that really gets me is you guys are all smart enough to now there's a hugh advantage to the the ondie memory controller and because of this AMD has an egde in performance.

Now it seems that AMD'S advantage here is having problems scaling up to higher memory speeds. So now Intel is going that route. Sometimes I get so confused.
Maybe intels cpus won't suffer this fate with the new Pressies but I am thinking the P3 arch. well also run into this wall.
So now Iam beginning to think Intel pressies may not be as bad as I precieved them to be . Man around and around we go. I well know much more with the release of the new Intel cpu's on 65 nano.chips if Intel gets the power usage and heat under controll netburst may still be alive Don't say it I know I know Ok! Lets say we can get to DDR2 to 1600 at low lati. And pressies can use that speed to its advantage. Who knows? Than indeed the ondie memory contoller well benefit the Pressies more than it will the conroe.

Confused yet I sure as hell am.
Click to expand...

AMD isn't having any problems scaling up in memory support. The JEDEC standard for DDR1 is DDR400 (PC3200 etc.). Any higher memory frequency would be non standard and AMD isn't going to officially release a processor with non-standard memory as a base. (Hidden overclocker options are just that, options)

DDR2 to 1600 is not going to happen. DDR 2 to 1000 maybe. DDR3 1600 isn't even on graphcis cards much less main system memory. AMD at least in the next revision, is going to DDR2 667 on die. Until the latencies of DDR2 667 and DDR 2 800 came (is still coming down) it was pointless for AMD to upgrade. There really isn't much difference in performance bewteen DDR 667 with higher latencies and DDR 400 with lower latencies. Now that DDR latencies and prices are coming down we notice AMD having the DDR 2 (and 3) cpus coming to market.

The world record is meant to be broken. 3.66 FX (MMMM drools)
Click to expand...

I was referring to AMDs new 500 standard and their does indeed seem to be a problem here. As some here know I am sure of that. I am not new to this so I will try to see it apples to apples ok
Click to expand...


Their new 500 DDR1 is not a standard. It is an overclocker option. The only official standard memory of the AMD64 is PC3200 (default, I am not referring to older cpus with 4 dimms at PC 2700 etc.) It would be like saying Intel is having problem scaling their DDR2 memory to DDR2 1000. It is an overclocking option not a standard.

In other words, try and find a AMD64 machine for sale with DDR1 250 memory modules out of the box without the disclaimer of non-standard memory usage.

Official Support
 
I

Intelia

Banned
May 12, 2005
832
0
0
Maybe I wasn't comperhending what Anandtech was saying here but that would be strange . Here's the link

http://www.anandtech.com/mb/showdoc.aspx?i=2457&p=2

I just reread this artical and it would seem Anandtech has revised their orginal artical or I read a differant artical were the test was conducted at DDR 500 either way there does seem to be a problem here . as this article would also indicate . The one I read orginally was tested and it did not look good.
 
I

imported_michaelpatrick33

Platinum Member
Jun 19, 2004
2,364
0
0
Originally posted by: Intelia
Maybe I wasn't comperhending what Anandtech was saying here but that would be strange . Here's the link

http://www.anandtech.com/mb/showdoc.aspx?i=2457&p=2

I just reread this artical and it would seem Anandtech has revised their orginal artical or I read a differant artical were the test was conducted at DDR 500 either way there does seem to be a problem here . as this article would also indicate . The one I read orginally was tested and it did not look good.
Click to expand...

I am not following you. What does that article have to do with memory? Yes the 233 and 250 memory dividers have been activated in the revision E processors but that is for overclockers (AMD even announced an official overclocker campaign with the FX-57 while keeping it unofficial, LOL). The PC3200 is still the standard memory for the AMD64 and FX and X2 on the 939 boards. The 4dimms at PC3200 is a new feature of the memory controller on the Manchester/San Diego/Toledo/Venice processors though it must be set at 2t timings. Of course real world testing done by Anandtech forum members only show about 3% average performance difference between 1t and 2t.
 
L

l Xes l

Banned
Feb 3, 2005
3,459
0
0
this isn't the "RECORD".. someone already broke 4ghz barrier w/ fx-55..
u should change the title ...
 
You must log in or register to reply here.

TRENDING THREADS

COMPANY
RESOURCES
FOLLOW
Top Bottom